Physical Therapy Use, Costs, and Value for Latent Classes of Good vs Poor Outcome in Patients Who Catastrophize About Their Pain Prior to Knee Arthroplasty.

This was a secondary analysis of trial data with all treatment arms combined, so it did not compare patients who received PT with patients who did not receive PT. The authors could not determine whether the poor outcome subgroup would have regressed without continued PT. The sample was selected for moderate to high pain catastrophizing, so results may not apply to all knee arthroplasty patients. PT visits and timing came from patient diaries, and cost estimates assumed 45-minute visits, CPT code 97110, and 20% patient cost-sharing, giving $115 per visit. Only 326 of 384 patients with knee arthroplasty completed at least 7 of 9 diaries, although age, sex and race did not differ significantly between included and excluded participants. In the sensitivity analysis that excluded 68 participants with additional orthopedic surgery, total PT cost differences were no longer significant, although cost per improvement patterns were similar. Good and poor outcome groups were latent classes with entropy 0.89 for WOMAC Pain and 0.87 for WOMAC Physical Function, and average posterior probabilities from 0.93 to 0.97.

The original KASTPain trial was funded by the National Institutes of Health, and the authors reported no competing interests.

Do not conclude that physical therapy should be stopped for patients with poor knee arthroplasty outcomes. This study did not compare patients who received PT with patients who did not receive PT, and the authors state they could not determine whether the poor outcome subgroup would have regressed without continued PT.